1. Field of the Invention
The present invention relates generally to an apparatus and method for controlling Automatic Repeat reQuest (ARQ) in a high-speed data transmission system, and in particular, to an apparatus and method for controlling Hybrid Automatic Repeat reQuest (HARQ) in a high-speed data transmission system.
2. Description of the Related Art
Mobile communication systems have been developed to provide a high-quality call service to mobile users. With the development of mobile communication systems, research is being conducted on a method for transmitting increasing amounts of data to users. In addition, mobile communication systems have evolved from an analog system to a digital system. Using the digital system, the mobile communication systems can now transmit increased amounts of data to users at higher speeds.
Generally, in digital mobile communication systems where a variation in channel condition is significant and different types of service traffic channels coexist with each other, a Hybrid Automatic Repeat reQuest (hereinafter referred to as “HARQ”) scheme is used to meet demand for high-speed data transmission, i.e., to increase transmission efficiency or transmission throughput. Particularly, as the commercialization of high-speed data transmission service has been realized, analysis and research are being actively conducted on technology for efficiently applying a HARQ scheme using error correction codes with a variable code rate, rather than a HARQ scheme using existing error correction codes with a fixed code rate. For a channel structure for high-speed transmission, a method of using high-level modulation such as 8-ary phase shift keying (8-PSK) and 16-ary quadrature amplitude modulation (16-QAM) in addition to the general binary phase shift keying (BPSK) or quadrature phase shift keying (QPSK), as a modulation scheme, is also taken into consideration.
Currently, a Code Division Multiple Access 2000 for Spreading Rate 1 (CDMA2000 1×) Evolution Data and Voice (EV-DV) system, which is a new transmission standard for a synchronous Third Generation Partnership Project (3GPP2) CDMA system, has adopted a coding scheme using quasi-complementary turbo codes (QCTC) as its standard. The quasi-complementary turbo codes provide a variable code rate to a coding scheme for a HARQ scheme over high-speed data and guarantee an improvement in soft combining performance based on HARQ. In the EV-DV system, packet data transmission/reception is performed by a HARQ or fast HARQ operation at a physical layer. This will be described in detail with reference to FIGS. 1 and 2.
FIG. 1 is a block diagram illustrating a relationship between an upper layer and a physical layer for ARQ processing according to the prior art. Referring to FIG. 1, a physical layer 110 decodes data received over a radio channel and makes decoded frame data. The physical layer 110 provides the decoded frame data to a medium access control (MAC) layer 120 called an upper layer. The MAC layer 120 determines whether the decoded frame data received from the physical layer 110 has a Protocol Data Unit (MuxPDU) error. When an error occurs, the MAC layer 120 retransmits the defective data. However, when no error occurs, the MAC layer 120 transmits a new frame. When such processing is performed in the MAC layer 120, since data decoded in the physical layer must be delivered to the upper layer to be processed, ARQ processing speed is undesirably decreased. In addition, since high-speed data processing must be performed, the load on the MAC layer 120 is increased. Hence, a structure in which part of an ARQ operation is performed in the physical layer as illustrated in FIG. 2 has been proposed.
FIG. 2 is a block diagram illustrating a relationship between an upper layer and a physical layer for improved fast (physical) HARQ processing. With reference to FIG. 2, a description will now be made of the relationship between an upper layer and a physical layer for improved fast HARQ processing.
In FIG. 2, part of an ARQ operation that has been performed in a MAC layer 230 is performed in a physical layer or its intermediate layer for fast ARQ response and processing. That is, in this scheme, a physical layer 200 has a basic physical layer 210 and a HARQ controller 220 performs the same operation as that of FIG. 1. The HARQ controller 220 performs part of the operation that has been performed in the conventional MAC layer. Therefore, the HARQ controller 220 is included in the physical layer in structure, but performs part of an operation of the MAC layer 230. Because of such characteristics, the HARQ controller 220 is often classified as a MAC layer. However, since the physical layer determines data retransmission, a processing time for the same data is shortened.
In addition, NAK transmission from the upper layer can disable soft combining on the same data, because the physical layer can maintain a soft-combined value for each symbol. However, since data symbols delivered from the physical layer to the MAC layer are all expressed with a binary value (0 or 1), although a symbol is repeated by retransmission, there is no way to soft combine the repeated symbol. The only method is to use a majority voting method for calculating the number of 0s and 1s for symbols having a binary value, and comparing the number of 0s with the number of 1s to determine a majority symbol. However, this method also cannot be used in the upper layer because of its required calculations.
In contrast, NAK transmission from the physical layer enables soft combining on code symbols for the same encoder packet, contributing to efficient utilization of channel resources. Therefore, it is preferable to dispose of the HARQ controller 220 under a multiplexing sublayer 230 of the MAC layer. That is, it is preferable for the MAC layer to perform an operation of the physical layer.
This structure has a fast processing time compared with a conventional ARQ control method operating based on a Radio Link Protocol (RLP). This method will now be compared with the existing method. In the conventional method of FIG. 1, a NAK signal is received through one packet transmission, and a round trip delay of a minimum of about 200 msec occurs up to a time when a retransmission packet is transmitted due to the NAK signal. On the contrary, in the method of FIG. 2, HARQ generates a very short round trip delay of a minimum of about several milliseconds. Therefore, it has a very good structure for implementing adaptive modulation and coding (AMC).
In order to actually operate HARQ with the structure of the upper layer and the physical layer of FIGS. 1 and 2, a HARQ transmission protocol of a transmitter for a retransmission request (i.e., NAK transmitted from a receiver) is required. For this, the 3GPP2 CDMA2000 1×EV-DV system uses Asynchronous and Adaptive Increment Redundancy (AAIR) for packet data transmission, and this will be described in brief below.
A base station asynchronously performs packet transmission to a corresponding mobile station according to the quality of a forward channel transmitted from the mobile station. At this point, a modulation scheme and a code rate of the transmission packet are adaptively applied according to a channel condition. In addition, a packet that failed to be transmitted during initial transmission is retransmitted. During retransmission, a code symbol pattern that is different from that used for the initial transmission can be transmitted. Such an AAIR HARQ scheme increases a signal-to-noise ratio (SNR) of packet data due to an increase in the number of retransmissions, and increases a coding gain due to a decrease in a code rate, thereby improving packet data transmission/reception performance.
A channel used for transmission of forward packet data in the 1×EV-DV system includes a forward packet data channel (F-PDCH) for payload traffic and a forward packet data control channel (F-PDCCH) for controlling the F-PDCH. F-PDCH is a channel for transmitting an encoder packet (EP) which is a transmission data block, and a maximum of up to 2 channels are used to simultaneously transmit their encoder packets to 2 mobile stations by using time division multiplexing (TDM)/code division multiplexing (CDM). An encoder packet is encoded by a turbo encoder, and then some of the encoded symbols are selected as a subpacket having a specific Increment. Redundancy (IR) pattern by OCTC symbol selection. The subpacket is a transmission unit for initial transmission and retransmission, and at each transmission, an IR pattern of a subpacket is identified by a subpacket identifier (SPID). A modulation scheme (QPSK, 8PSK or 16QAM) and a transmission slot length (1, 2 or 4 slots) of the subpacket are determined according to forward channel quality information transmitted from a mobile station and resources (e.g., the number of Walsh codes and power assignable to F-PDCH) of a base station.
Information related to demodulation and decoding of F-PDCH is multiplexed with F-PDCH through other orthogonal channels for the same slot period, and then transmitted over the F-PDCCH which is a control channel. Information included in the F-PDCCH is very important in performing a physical layer HARQ operation by a mobile station, and requires the following:
1) fragmented Walsh code information available for F-PDCH every several tens to several hundreds of milliseconds;
2) MAC_ID: MAC_ID of a mobile station (MS) to which F-PDCH is assigned;
3) ACID: ID for identifying 4 ARQ channels (ARQ channel ID);
4) SPID: ID for identifying an IR pattern of a subpacket;
5) EP_NEW: information for distinguishing two consecutive encoder packets in the same ARQ channel;
6) EP_SIZE: a bit size of an encoder packet; and
7) Last Walsh Code Index (LWCI): information on a Walsh code used for F-PDCH.
Meanwhile, packet data reception in a mobile station is performed from the decoding of F-PDCCH. A mobile station first decodes F-PDCCH to determine whether its own packet is being transmitted, and if it is determined that the transmitted packet is its own packet, the mobile station performs demodulation and decoding on F-PDCH. If a currently received subpacket is a subpacket that was retransmitted for a previously received encoder packet, the mobile station performs decoding after code-combining the currently received subpacket with code symbols of an encoder packet that was previously received and stored therein. If the decoding is successful, the mobile station transmits an ACK signal over a reverse ACK/NAK transmission channel (R-ACKCH), allowing the base station to transmit a subpacket for the next encoder packet. If the decoding is not successful, the mobile station transmits a NAK signal, requesting the base station to transmit a subpacket for the same encoder packet.
A unit for which a physical layer's HARQ operation is performed on one encoder packet is called an “ARQ channel.” In the CDMA2000 1×EV-DV system, a maximum of 4 ARQ channels can simultaneously operate, and these are called “N=4 fast HARQ channels.”
In the 1×EV-DV standard, it is provided that ACK/NAK Delay which is required to be performed by a mobile station a packet reception operation and transmitting ACK/NAK and the number of simultaneously available ARQ 10 channels should be provided to a base station by the mobile station, and this becomes an implementation issue for a mobile station. Therefore, a possible ACK/NAK Delay supported by the mobile station is 1 slot (=1.25 msec) or 2 slots (2.5 msec), and the possible number of ARQ channels is 2, 3 or 4. With reference to FIGS. 3 and 4, a description will now be made of an operation depending on ACK/NAK Delay and the number of ARQ channels.
FIG. 3 is a timing diagram illustrating an example of a timing relationship between a base station and a mobile station for ACK/NAK Delay=1 slot during a HARQ operation in a mobile communication system, and FIG. 4 is a timing diagram illustrating an example of a timing relationship between a base station and a mobile station for ACK/NAK Delay=2 slots during a HARQ operation in a mobile communication system.
It will be assumed in FIGS. 3 and 4 that a F-PDCH is assigned to a mobile station A. In addition, for the convenience of explanation, indexes are sequentially assigned to time slots of both a base station (BS) and a mobile station (MS) from a 0th time slot beginning at a particular time. Further, in FIGS. 3 and 4, A(x,y) has the following meaning. Hatched parts denote data to be transmitted to the mobile station A. In addition, ‘x’ denotes an ARQ channel, and ‘y’ denotes an index for distinguishing an IR pattern for the same encoder packet.
Based on this, a description will now be made of FIG. 3 in which ACK/NAK Delay is 1 slot.
Referring to FIG. 3, data from a base station is transmitted to a mobile station A at a 0th slot. Then, the mobile station A receives the packet data at the same slot. In FIGS. 3 and 4, the base station and the mobile station have different slot start points due to transmission delay occurring between the mobile station and the base station on the basis of an absolute time. At this point, the base station transmits packet data and a packet data control signal over a F-PDCH and a F-PDCCH, respectively. Then, the mobile station A determines whether the data has an error, for a one-slot processing time, and thereafter, transmits ACK or NAK to the base station. The “processing time” refers to a time required for performing demodulation and decoding on received packet data for one slot, and transmitting the result at the next slot over a reverse channel (R-ACKCH). For example, in FIG. 3, NAK is transmitted. The base station then receives the NAK at a 3rd slot, and schedules retransmission of the defective data at a 4th slot.
Thereafter, the base station transmits data of a different pattern for the same encoder packet according to the scheduling result.
Next, a description will be made of FIG. 4 in which ACK/NAK Delay is 2 slots. It will be assumed in FIG. 4 that an error has occurred in a first data packet among the data packets transmitted from a base station to a mobile station A, and the description focus on the first data packet. Since the delay time is 2 slots, the base station continuously transmits packet data to the mobile station A at a 0th slot, a 1st slot and a 2nd slot. The mobile station then checks an error of the data transmitted at the 0th slot for a period between the 1st slot and 2nd slot, checks an error of the data transmitted at the 1st slot for a period between the 2nd slot and 3rd slot, and checks an error of the data transmitted at the 2nd slot for a period between the 3rd slot and 4th slot. ACK/NAK for the data received at the 0th slot is transmitted at the 3rd slot, ACK/NAK for the data received at the 1st slot is transmitted at the 4th slot, and ACK/NAK for the data received at the 2nd slot is transmitted at a 5th slot. If the base station receives, at the 4th slot, NAK for the packet data transmitted at the 0th slot, the base station performs, at the next slot, retransmission on an encoder packet transmitted at the 0th slot. The retransmitted packet data is the same packet as the previously transmitted packet but has a different IR pattern.
As can be understood from FIGS. 3 and 4, the mobile station performs synchronous ACK/NAK transmission in which the mobile station must transmit ACK or NAK for a received packet after a lapse of 1 slot or 2 slots. The base station performs asynchronous ACK/NAK transmission in which the base station can transmit a packet at any slot after receiving ACK/NAK for a packet previously transmitted by the mobile station for the same ARQ channel.
In addition, FIGS. 3 and 4 illustrate a 1-channel ARQ operation and a 4-channel ARQ operation, respectively. In the 1-channel ARQ operation of FIG. 3, data transmission to one mobile station uses only a part of base station resources, decreasing a packet data rate of a corresponding mobile station. In contrast, in the 4-channel ARQ operation of FIG. 4, one mobile station can use the entire resources of the base station, so a corresponding mobile station can obtain a maximum packet data rate.
As described above, it is possible to enable fast ARQ response and processing by moving ARQ control that was achieved in the conventional upper layer, under a multiplexing layer. However, this is a mere logical resolution in the standard, and the following problems arise in its actual implementation.
First, currently, most systems implement an upper layer including a multiplexing layer by software loaded in a central processing unit (CPU). However, in the case of a mobile station, its CPU does not have high processing speed and capability. Therefore, in implementing a HARQ protocol requiring a fast response in the CPU, overload may occur in a clock of the CPU. As a result, the mobile station may fail to perform its normal operation. Particularly, such a problem serves as a big hindrance to implementation, when power consumption of the mobile station is an implementation limitation factor of the system.
Second, transmission interrupt of coded data and a processing delay due to the interrupt, which impose an overload to the CPU, must be reduced to process high-speed transmission data. Therefore, consideration should be taken of a method for reducing data processing interrupts which may occur every 1.25 msec.
Third, in order to support N-channel HARQ, N independent HARQ controllers are required. Therefore, if N is increased, the number of HARQ controllers is also increased, resulting in increased power consumption and complexity. Thus, in implementation, the number of HARQ controllers must be minimized.
Fourth, in order to support N-channel HARQ, N independent turbo decoders are required. Therefore, if N is increased, the number of turbo decoders is also increased, resulting in increased in power consumption and complexity. Thus, in implementation, the number of turbo decoders must be minimized.
Fifth, in the standard, ACK_DELAY=1 slot and ACK_DELAY=2 slots described in conjunction with FIGS. 3 and 4 are exclusive options. However, in implementing a mobile station, a structure for varying an operating clock of the mobile station by selectively demultiplying/multiplying the operating clock for low power consumption is considered, so a mobile station's structure capable of applying both ACK_DELAYs in one mobile station must be designed.
Sixth, unlike the conventional data traffic, an encoder packet which is a data block transmitted over a forward packet data channel (F-PDCH) can change its transmission scheme every 1.25 msec. Therefore, a new structure for transmitting, every 1.25 msec, channel structure information that is transmitted once during setup of a data channel is required.
Finally, the other control information necessary for application of a mobile station is transmitted by a base station over a forward packet data control channel (F-PDCCH) which is a traffic control channel. Therefore, the mobile station must efficiently perform an operation of detecting the control information and delivering the detected control information to the upper layer within a short time.